SC3 is a user-friendly tool for unsupervised clustering of single-cell RNA-seq (scRNA-seq) data, combining multiple clustering solutions through a consensus approach to achieve high accuracy and robustness. It identifies subclones based on transcriptomes from neoplastic cells. SC3 is an R-package that integrates with Bioconductor and scater, making it easy to incorporate into bioinformatic workflows. The SC3 pipeline involves gene filtering, distance calculations, transformations, k-means clustering, and consensus clustering. The consensus matrix is used to determine the final clustering through complete-linkage hierarchical clustering. SC3 was tested on six publicly available datasets, showing high accuracy and stability. It outperformed other methods like tSNE+kmeans, pcaReduce, SNN-Cliq, SINCERA, and SEURAT in terms of accuracy and stability. SC3 can identify differentially expressed genes, marker genes, and outlier cells. It was also used to analyze a large Drop-Seq dataset with 44,808 cells, showing good agreement with the original authors. SC3 can identify subclones across patients, as demonstrated by analyzing two patients with myeloproliferative neoplasms. The tool uses a hybrid approach combining unsupervised and supervised methodologies to handle larger datasets. SC3 also provides a method based on Random Matrix Theory to determine the number of clusters. The tool is highly stable and can handle rare cell types, though it may fail to identify very rare cell types. SC3 is benchmarked against other methods using the Adjusted Rand Index (ARI) to measure similarity between clusterings. The tool is suitable for both basic biology and clinical applications, providing insights into cell-type characterization and subclone identification.SC3 is a user-friendly tool for unsupervised clustering of single-cell RNA-seq (scRNA-seq) data, combining multiple clustering solutions through a consensus approach to achieve high accuracy and robustness. It identifies subclones based on transcriptomes from neoplastic cells. SC3 is an R-package that integrates with Bioconductor and scater, making it easy to incorporate into bioinformatic workflows. The SC3 pipeline involves gene filtering, distance calculations, transformations, k-means clustering, and consensus clustering. The consensus matrix is used to determine the final clustering through complete-linkage hierarchical clustering. SC3 was tested on six publicly available datasets, showing high accuracy and stability. It outperformed other methods like tSNE+kmeans, pcaReduce, SNN-Cliq, SINCERA, and SEURAT in terms of accuracy and stability. SC3 can identify differentially expressed genes, marker genes, and outlier cells. It was also used to analyze a large Drop-Seq dataset with 44,808 cells, showing good agreement with the original authors. SC3 can identify subclones across patients, as demonstrated by analyzing two patients with myeloproliferative neoplasms. The tool uses a hybrid approach combining unsupervised and supervised methodologies to handle larger datasets. SC3 also provides a method based on Random Matrix Theory to determine the number of clusters. The tool is highly stable and can handle rare cell types, though it may fail to identify very rare cell types. SC3 is benchmarked against other methods using the Adjusted Rand Index (ARI) to measure similarity between clusterings. The tool is suitable for both basic biology and clinical applications, providing insights into cell-type characterization and subclone identification.